Full Text

Int. J. Dev. BioI. 42: 763-774 (1998)
Original
Article
Evidence for non-axial AlP patterning in the nonneural
ectoderm of Xenopus and zebrafish pregastrula embryos
E. MARY READ, ADAM R.F. RODAWAY, BYRON NEAVE', NICK BRANDON', NIGEL HOLDEW,
ROGER K. PATIENT and MAGGIE E. WALMSLEY'
Developmental
Biology Research
Centre, King's College London,
The Randal/Institute,
London,
United Kingdom
ABSTRACT Recent studies in early Xenopus and zebrafish embryos have demonstrated
that
posteriorizing,
non-axial signals arising from outside the organizer (or shield) contribute to AlP
patterning of the neural axis, in contradiction tathe classical Spemann model in which such signals
were proposed to be solely organizer derived. OUfstudies on the early expression of the transcription factors GATA-2 and 3 in both Xenopus and zebrafish nonneural ectoderm lend support to the
existence of such non-axial signaling in the AlP axis. Thus we find that the earliest expression of
GATA-2 and 3 is located in nonneural ectoderm and is strongly patterned in a graded manner along
the AlP axis, being high anteriorly and absent from the most posterior regions. This results by early
neurula stages in three broad zones: an anterior region which is positive for both GATA-2 and 3, a
middle region which is positive for GATA-2 alone and a posterior region in which neither gene is
expressed. These regions correspond
to head, trunk and tail ectoderm and may represent the
beginnings of functional segmentation
of nonneural ectoderm, as suggested in the concept of the
'ectomere'.
We find that AlP patterning of GATA expression in nonneural ectoderm may occur as
early as late blastulalearly
gastrula stages. We investigate which posteriorizing
signals might
contribute to such distinct non axial ectodermal patterning in the AlP axis and provide evidencethat
both FGF and a Wnt family member contribute towards the final AlP pattern of GATA expression
in nonneural ectoderm.
KEY WORDS: A/I' patterning, nanneural ectodam, GA TAfacton,
Introduction
The induction and patterning of the embryonic nervous system
has been the focus of much experimental effort over a number of
years [for reviews see (Doniach, 1992; Lumsden and Krumlauf,
1996)]. In contrast, very little is known aboul the formation and
patterning of nonneural ectoderm. Recent experiments in early
Xenopus and zebrafish embryos reveal the activity of posteriorizing,
non-axial signals arising from outside the organizer (or shield) in
contradiction to the classical Spemann model in which such signals
were proposed to be solely organizer derived (Bang ef al., 1997;
Woo and Fraser, 1997). Clearly such radial NP signals should
operate within nonneural as well as neural ectoderm, but, other
than isolated reports of the expression in nonneural ectoderm of
single hoxgenes (Condie and Harland, 1987; Kolmand Sive, 1995;
von Bubnoff ef al., 1995), patterned gene expression within
nonneural ectoderm has not been reported.
Xenopus, ze&rajish, FGF, Wnt
The importance of the non neural ectoderm in early development is indicated by its role in a number of inductive epithelialmesenchymal interactions which are required for early embryonic
patterning. Thus ectodermal signaling to mesenchyme is important
for initial patterning events in tooth development
in urodeles
(Lumsden, 1988), forthe directional guidance of neural crest and
migrating pronephric duct in axolotl (Lofberg ef al., 1985; Drawbridge ef ai, 1995), and for the stimulation of hematopoietic
differentiation within the blood islands in Xenopus (Maeno et al.,
1996).ln earlycranio-facial
development, matrix mediated inleractions between epithelia and mesenchyme, operating locally, are
A,bbreviations used in this paper: A/P, amcro/posterior;
D/V, dorsal/ventral;
nnc, non neural extoderm;
bFGF, basic fibroblast
gro\'..th factor; eFGF,
embryonic
fibroblast
growth
factor;
RA, retinoic
acid; EMP, bone
morphogenic
protein;
MET, mid blastula transition;
TGF, transforming
growth factor.
*Address
Kingdom.
for reprints:
Developmental
Biology Research Centre, King's College London, The Randall Institute,
Fax; 0171 4979078. e-mail: [email protected]
tpresent
'Present
+Present
address:
address:
address:
0214-6282/98/$10.00
()
UBC Pre~~
Prinled in Spain
26-29, Drury Lane, London WC2B 5RL, United
Department of Biomedical Sciences, University of Sheffield, Western Bank, Sheffield, Yorkshire, S10 2TN, United Kingdom.
MRC Laboratory of Molecular Cell Biology, University College, Gower Street, London WC1 5BT, United Kingdom.
Department of Anatomy and Developmental
Biology, University College, Gower Street, London, WC1E 6BT, United Kingdom.
764
E. M. Read ef al.
Krox 20
0lx2
a
c
b
d
thought to playa role in directing head mesenchyme differentiation into cartilage and bone,
thus specitying skeletal pattern (Thor0900d,
1988). In addition, instructive interactions from
nonneural to neural ectoderm have been demonstrated in the induction of dorsal neural tube
cell types including neural crest (Dickinson et
Slug
Hox B9
N
;!;
<0:
CJ
al., 1995;
co
;!;
<0:
CJ
d
k
nolo
o
d
noto
"E"'OPUSiZUUf'SH""S'~""'"
-l.
.
'
,~,
v
.""TH ""'
g..,,,,,_
j .~
Fig. 1. Conserved
boundaries
of expression
of GATA factors in the nonneural
ectoderm
of frog and fish gastrulae.
Doubte whole-mount
in situ hybridizations of
Xenopus early neurula embryos stained for GA TA-2 (a,c,e,g) or GA TA-3 (b,d,f,h) with the
anterior neural marker, Otx2 (a and bJ, the hindbrain marker Krox20 (c and d), the spmal
cord marker HoxB9 (e and f) and neural crest marker, slug (g and h). Anterror ISto the left
and posterior to the right except for a and b which are anterror views, (a) Embryo stained
for Otx2 (turquoise) and GA TA-2 (purple). At the most anterior end GA TA-2 is expressed
in the cement gland and abuts Otx2 whereas more posteriorly a gap can be seen between
GA TA-2 and Otx2 (asterisk). (bJ Double staining for GA TA-3 (purple) and Otx2 (turquoise).
At the most anterior end GA TA-3 expressed in the cement gland overlaps the expression
of Otx2. More posteriorly, a gap exists between GA TA-3 and Otx2 expression (asterisk).
(e) Dorsal view of an embryo stained for GA TA-2 (purple) and Krox 20 (turquoise) marking
rhombomeres
3 and 5 of the hindbrain (r3 and (5). A gap can be seen between GA TA-2 and
Krox20 expression in r5 (arrow). (d) Dorsal view of an embryo stained for GA TA-3 (purple)
and Krox20 (turquoise). A gap is seen between GA TA-3 and Krox20 staining In r5 (arrow).
(e and 1) Lateral view of embryos stained for GATA-2 and 3 (purple) and Hox89 (89,
turquoise). Note that GA TA-2Is expressed more posteriorly than GA TA-3. (9 andh) Dorsal
views of embryos stained for GA TA-2 (purple, g) or GA TA-3 (purple, h) and the neural crest
marker slug (turquoise, g and h). Note the gap between
GA TA and slug expression
(astensk, g and h). Ii and j) Double whole-mount
in Situ hybridization of Xbra (turquoise,
i and j) with GATA-2 (dark blue, i) and GATA-3 (dark blue, J) In Xenopus mid-gastrula
embryos (the notocordal Xbra signal is not yet visible, stage 11). Note that GA TA-2 abuts
Xbra expression, (i), but a gap is seen between GA TA-3 and Xbra expression, (arrow, j). yp,
yolk plug; d, dorsal; v, ventral; anterior is to the left. (k) A 70% epiboly (gastrula) zebrafish
embryo, stained for gta2 (blue) and ntl (red). gta2 expression abuts ntl at all gastrula
stages. (I) An intact 70-80% (gastrula) zebra fish embryo stained for gta3 (blue) and ntl
(red). Note the gap between gta3 and ntl (arrow), This gap is also seen at eartier gastrula
stages (data not shown). At late gastrula, ntl is expressed in the notochord (noto). (m) 17
mm saggital section through the region marked by the arrow in ishowing the gap (arrow)
between gta3 (blue) and ntl (red) expression. a, anterior; p, postenor; v, ventral. (n) A flat
mount of a zebrafish embryo stained for gta3 (blue) and dlx3 (expressed
strongly in
pIa codal tissue at this stage) and pax2 (a marker of the midbrain/hindbrain border) both
stained in red. Note the gap between dlx3 and pax2 staining representing neural crest. a,
anterior; p, posterior; v, ventral. (0) Diagrammatic representation
ofaXenopus/zebrafish
gastrula embryo showing the boundaries of expression of GA TA-2 and 3 between neural
and nonneural ectoderm and between nonneural ectoderm and mesoderm
which are
conserved between the two species.
Liem
ef al., 1995).
Such
diversity
of
function implies that the nonneural ectoderm
cannot exist as a blank sheet but must be
patterned in order that a variety of regional
inductive, instructive or permissive interactions
can occur. Indeed support for the concept of
patterning in the nonneural ectoderm came from
studies involving transplantation of anterior neural crest and associated superficial nonneural
ectoderm between quail and chick neurula embryos. These experiments led to the suggestion
that not only is the neural tube segmentally
divided into units (the neuromeres) along the N
P axis but so too is neural crest and nonneural
ectoderm, thus defining an early developmental
unit -the 'ectomere' (Couly and Le Douarin,
1990).
The identity of nonneural ectoderm is dependent on BMP-4. In the dorsoventral (ON)
axis, BMP-4 promotes epidermal development
and inhibits neural tissue formation (Graff et a/.,
1994; Sasai ef al., 1995; Wilson and HemmatiBrivanlou, 1995; Neave ef al., 1997). Inhibition
of BMP receptor activity by injection of a dominant negative BMP2/4 receptor into ventral regions of the Xenopus embryo results in formation of a second neural axis (Graff el al., 1994;
Suzuki el al., 1994; Schmidt ef al., 1995). The
same effect is produced by ventral injections of
RNA coding for cleavage mutant BMP2, 4 and 7
proteins (Hawley ef al., 1995). In addition,
neuralization of dissociated animal cap ectoder.
mal cells is now thought to be due to dilution of
epidermal promoting BMp.4 in the animal cap
(Wilson and Hemmati-Brivanlou,
1995). Inhibition of BMPA receptor activity in animal caps
leads to formation of neural tissue and BMP-4
expressed during gastrulation blocks neural induction by noggin and follistatin (Sasai ef al.,
1995). More recently, both nog9in and chordin
have been shown to antagonize BMP-4 by direct interaction between the proteins, preventing binding of BMP-4 to its receptor (Piccolo ef
al., 1996; Zimmerman ef al., 1996). Thus BMP4 activity is an essential
requirement
for
nonneural ectoderm identity and inhibition of
BMP-4 activity is necessary for neural tissue
formation.
In the animal/vegetal
axis of late blastula!
early gastrula stages (which for ectoderm is
rou9hly equivaient to the NP axis). prospective
non neural ectoderm is in apposition to the ring
of mesendoderm induced around the equator of
FGF alld Wilts patlem lIolllleural ectoderm
the embryo. So the signaling molecules likely to define the limits of
nonneural ectoderm in this axis are the mesoderm inducing growth
factors belonging to the FGF and TGFp families (Smith, 1989). In
addition, recent studies in Xenopus from a number of labs have
demonstrated that, during gastrula stages, bFGF is capable of
inducing neural tissue directly, independent of mesoderm
induction (Kengaku and Okamoto, 1993; Kengaku and Okamoto, 1995;
Lamband Harland, 1995; Launay et al., 1996), and of posteriori zing
anterior neural tissue induced by noggin (Lamb and Harland,
1995), follistatin or a dominant negative activin receptor, 61 XAR 1
(Cox and Hemmati-Brivanlou,
1995). Furthermore, a dose-dependent effect of bFGF on neural posteriorization has been demonstrated (Kengaku and Okamoto, 1995). In Xenopus, FGF is
enriched in posterior/vegetal regions of the gastrula embryo in and
around the blastopore (Isaacs et al., 1994) and could therefore
posteriorize nonneural as well as neural ectoderm.
Retinoic acid (RA) has also been demonstrated to act as a
posteriorizing molecule during gastrula stages. Addition of exogenous RAduring blastula and gastrula stages results in concentration-dependent disruptions of the AlP axis, affecting both mesodermal and ectodermal tissues (Dursten et al., 1989; Sive el al., 1990;
Ruiz i Altaba and Jessell, 1991). Molecular analysis shows that RA
treatment can lead to a concentration-dependent
decrease in
some anteriorly expressed genes and increase of some posterior
genes (Sive et al., 1990; Papalopulu and Kintner, 1996). Therefore,
although reports on the location of retinoid activity in the embryo
are variable, RA is another candidate for posteriorization during the
blastula and gastrula stages (Dursten et al., 1989; Chen et al.,
1994; Creech Krah et al., 1994).
Members of the Wnt family have also been shown to have
posteriorizing activities within Xenopus neural ectoderm. Although
incapable of inducing neural tissue per se, both XWnt3A RNA and
XWnt8 in a DNA construct which delays expression until after MBT
are capable of posteriorizing anterior neural tissue (McGrew et al.,
1995; Fredieu et al., 1997). Very recently XWnt3A was reported to
be expressed in posterior mesoderm and ectoderm during gastrula
and neurula stages (McGrew el al., 1997), and at least one other
member of the Wnt family -XWnt8 is also expressed in posterior
non-axial mesoderm during gastrulation (Christian and Moon,
1993). Therefore, like FGF, these Wntscould posteriorize nonneural
as well as neural ectoderm during the period of early neural
induction.
The Wnts, therefore,
represent
a third possible
posteriorizing activity which might pattern the nonneural ectoderm
during gastrulation.
Expression of the genes involved in early ectomesenchymal
signaling activities will be driven by transcription factors located in
the ectoderm at the appropriate time in development. Although the
transcription factors GAT A-2 and 3 are more commonly associated
with blood, we have shown that the earliest expression of these
factors in Xenopus and zebrafish embryos is primarily in ventral
ectoderm rather than mesoderm. This GAT A factor expression in
the ectoderm is transitory starting before the onset of gastrulation
and declining during neurula stages (Walmsley et al., 1994; Neave
et al., 1995; Bertwistle et al., 1996). The importance of GATA
factors for early ectodermal identity is clearly indicated by mutations in the Drosophifa and C. elegans GAT A homologs, pannier
and EL T-t (Ramain et al., 1993; Page et al., 1997). In this study,
using positionally defined neural, placodal and mesodermal markers, we demonstrate
that GAT A-2 and 3 predominate in the
non neural ectoderm of both frogs and fish during late blastula,
NCAM
765
__
GATA-3 __
- - .... ..
GATA-2__
EF1-a
--
...........
ui d50 tBR ui d50 tBR ui d50 tBR .ANA
I
I I
I I
I
4 hours
5.5 hours
0 IN
Fig. 2. Expression ofGATA.2 and 3 in nonneural ectoderm is depend.
ent on BMP-4 signaling. lng of a dominant negative truncated 8MP
receptor RNA (lanes tBR) or of a control RNA fd5OJ. was injected in to the
animal pole of all four bfastomeres of 4--cell stage embryos. Animal caps
were removed at srage 8 along wirh uninjecred caps (ui fanes). and cultured
for 4 h. 5.5 h or overnighr to stage 14 (OIN Janes). RNA prepared from caps
was probed by ribonuclease protection for the expression of GA TA-2 and
3, NCAM and EF1-a as a loading control.
gastrula and early neurula stages, and that the boundaries of
expression are conserved between the two species and may be
defined by the same signaling molecules. We show that GAT A
expression is patterned in a graded manner in the AlP axis being
high anteriorly and absent from posterior regions resulting in three
broad zones corresponding to head. trunk and tail ectoderm. This
distinctive AlP paherning of the GAT A factors does not appear to
result from RA activity. We present evidence that, as reported for
neural ectoderm, FGF and XWnt3A can contribute to this graded
pattern of expression in nonneural ectoderm, suggesting that
these molecules playa role in patterning all ectoderm and represent at least part of the non-axial signaling which patterns pregastrula embryos in the AlP axis.
Results
GA TA-2 and 3 expression in Xenopus ectoderm is patterned
along the DN alld AlP axes
In the DN axis at neurula stages, both GAT A-2 and 3 are
expressed in the sensorial layer of the ectoderm from the ventral
mid line to a point very close to the edge of the neural plate (Fig. 1e
and f). This indicates that, at this stage, both factors may be
expressed only in the nonneural ectoderm. The separation of the
GAT A-2 and 3 domains from the neural plate is demonstrated by
a small gap that can be seen between the limits of both GATA
766
E. M. Read et al.
3 and Krox20 (Fig. 1c and d) represents both
early neural crest and placodal precursors.
This was also demonstrated in zebrafish. Figure 1n shows a flat mount of a zebrafish
embryo at 95-1 00% epiboly probing for gta3in
purple, marking ventral ectoderm, pax2in red,
marking neural ectoderm (Krauss elal., 1991),
and dlx3 in red, which at this stage is strongly
expressed in presumptive
placodal tissue
(Akimenko el al., 1994). As in Xenopus, gla3
expression
in the ON axis does not abut the
Xenopus GATA-2
neural plate. It does, however, abut the dlx3
expressing placodal precursors. The gap between dlx3 and the neural plate marker pax2
represents
neural crest precursors.
Thus
I
GAT A expression is confined to the nonneuraJ
ectoderm.
In the NP axis, expression of GAT A-2 and
3 are both patterned in a rostra-caudal gradient with GATA-2 being expressed more
'd
Xenopus GATA-3
posteriorly than GATA-3 (Fig. 1c.h). In order
to define the boundaries of expression of
GATA-2 and 3 in non neural ectoderm within
this axis, early neurula Xenopus embryos
were examined by double whole-mount in situ
hybridization, probing for the GATA factors
and the neural markers Olx2 (defining anterior neurectoderm
-Pannese
et a/., 1995),
Krox20(expressed
in hindbrain rhombomeres
d
3 and 5 -Bradley
el a/., 1992) and HoxB9
which is expressed in the spinal cord (Wright
a
a.
el a/., 1990; Godsave et al., 1994). Both
Con
GAT A-2 and 3 are strongly expressed at the
most anterior edge of the nonneural ectoderm
as defined by the position of Otx2expression
(Fig. 1aand b) and their expression decreases
v
v
d
d
gradually in a posterior direction (Fig. 1c-f).
Both GAT A factors are also expressed in the
cement gland overlying the ventral forebrain
p
with GAT A-3 being expressed more dorsally
Zebra/ish qta3
then GATA-2 (Fig. 1b). Lateral views demonstrate that, for both GAT A factors, expression
Fig. 3. FGF represses the expression of GATA-2 and 3 (gta2 and 3) in the non neural ectoderm
falls away from the dorsal midline resulting in
of frog and fish embryos and restricts expression to regions outside the germ ring. Xenopus
their expression extending further posteriorly
(a-d) and zebrafish (e-j) embryos were either untreated (a,c,e.g and j) or injected with 2pg of eFGF
on the ventral side of the embryo (Fig. 1e and
RNA at the single cell stage for Xenopus (b and d) or with 220 pg eFGF RNA (note that this RNA
f). The ventral posterior limit ot GATA-3 exwas not made with the Megascript kit and was capped less successfulfy, thus more was needed
to produce an effect) at the 1-4 cell stage for zebra fish (f and h) or with 40pg XFD RNA at the 1pression
lies
somewhere
between
4 cell stage (i). Embryos were examined by whole-mount in situ hybridization staining for GA TArhombomere 5 in the hindbrain, as defined by
2 (a and b), GA TA-3 (c and d), gta2 (e and f) and gta3 (g and h). Note that there ISno gta2 staining
Krox20 (Fig. 1d), and the anterior limit of the
in f): the background is caused by yolk which has been fast from the embryo in e). (i) Zebra fish
spinal cord defined by HoxB9 staining (Fig.
gastrula embryo injected with XFD RNA and starned for gta2 (blue) and ntl (red), Note that gta2
1f), whereas GAT A-2 expression extends more
expression reaches down to the involuting edge on the ventral side where ntl has been eliminated.
posteriorly to a position well down the spinal
cord (Fig. 1e). Dorsally, near the boundary of
factors and Krox20expression
(Bradley et a/., 1992) in rhombomere
the neural plate, GATA-3 expression falls off just posterior to
5 (Fig. 1c and d, arrow). To define more closely the gap between
rhombomere
5 (Fig. 1d and f), whereas GATA-2 expression
neural and nonneural ectoderm in anterior regions, embryos were
reaches as far as the anterior limit of HoxB9 staining, the junction
stained for slug, a marker of neural crest (Mayor el al., 1995), along
between hindbrain and spinal cord (Fig. 1c and e).
with GATA-2 and 3. In this case, we could still detect a gap
In summary, GATA.2 and 3 are expressed in an antero-poste.
(asterisk, Fig. 1g and h) between GAT A positive ectoderm and slug
rior gradient in non neural ectoderm, high in the anterior and low or
positive neural crest. Therefore the gap seen between GAT A.2 and
absent in the more posterior regions of the embryo. Both dorsally
untreated
eFGF
,
f
v
9
j
h
FCF and Wnts pattern nonnel/ral ectoderm
12
Stage
RA
n
-
13
n
n
- .- -
+
in sitl/
18
.
- +
+
----GATA-3
. .... .
.
-
EF1-a
..~~,!
~
--I~
t;
,-
----GATA-2
*.
2
Keratin
. . . .1
____ EF1-a
.--
3
Fig. 4. RA is not responsible
4
5
6
for patterning
7
GATA-2 and 3 in nonneural
were isolated at stage 8-9, and treated with lCt7M
RA until stage 11. Caps were then transferred and aI/owed to develop
ectoderm.
Animal caps
without RA RNase protection was carried out on treated caps (+Jand on
untreated controls (-)taken at stages 12, 13 and 18. In this experiment there
is little effect on GATA-2 but, GATA-3is increased in RA treated caps at all
stages.
Inthe same experiment epidermal keratin is decreased.
EF1-a
was
for GAT
A-2
or 3 and
the
mesodermal
marker
Xbra
767
(Smith
et aI., 1991). At these early stages, within the ectodermal
layer,
the animal/vegetal
axis is roughly equivalent to the future AlP
axis. A lateral view of such embryos shows that, even during early
gastrulation, GATA-2 is expressed more posteriorly than GAT A3 (Fig. 1i and D. Whereas ventrally, the limit ot GATA-2 extends
posteriorly to touch the boundary of expression of Xbra, GAT A-3
expression falls short of it (arrow, Fig. 1j). In addition, the expression of both GATA-2
and 3 has already been restricted dorsally.
Thus, in Xenopus, the localization of GATA expression to the
nonneural ectoderm is established during gastrulation. Completion of the gastrulation movements will lead to the pattern seen in
neurula embryos. In the zebrafish, similar boundaries of expression can be seen during gastrulation
(Fig. 1 k and I). A gastrula
stage embryo
probed for gta2 and nt! (the zebrafish
homolog of
Xbra, (Schulte-Merker
et aI., 1994) demonstrates
that gta2 abuts
nt! ventrally
(Fig. 1 k), whereas
there is a clear gap between
the
ventral boundaries of nt!and gta3(Fig.
11, arrow). In the DN axis,
gta2 and 3 have been restricted
dorsally
and expression
is seen
only in ventral
nonneural
ectoderm
(Fig. 1 k and I). A saggital
section of the embryo shown in Figure 11clearly demonstrates the
gap between
gta3 and nt! expression
(arrow, Fig. 1m).
If we combine the data from frogs and fish, we may define the
expression domains of GATA-2 and 3 (gta2 and 3) as nonneural
A
used as loading control and tRNA was probed as a negative control (lane
7). The down regulation of epidermal keratin probably representsan effect
of the RA on differentiation of the ectoderm.
. ..-_
.----
and ventrally, GAT A-2 expression extends more posteriorly than
GAT A-3. Both factors are expressed strongly ventrally and extend
laterally to outline the anterior neural plate, with a small gap
between the boundaries of neural and non neural ectoderm which
may represent neural crest and placodal tissue. The distance from
the dorsal midline of GAT A-2 and 3 expression increases considerably towards the posterior end of the embryo, where the neural
plate border is less well defined, and expression extends more
posteriorly on the ventral side.
To understand how these patterns of expression in neurula
embryos might have been established we looked at mid to late
gastrula stage Xenopus embryos probed by double whole-mount
.
.. ""
GATA-3
"-
....... . _
.,
.,
2
Xbra
3
4
5
6
7
8
GATA-2
EF1-a
9
B
Fig. 5. FGF posteriorizes
blastula nonneural ectoderm and down
regulates GATA.2 and 3 in the absence of mesoderm
and neural
induction. (AJ GATA-2 and 3 are down regulated in the absence of and at
low levels of mesoderm induction. eFGF RNA was Injected in to the animal
pole of single cell stage Xenopus embryos over a range of 60 fg to 2.5 ng
(lanes 1 to 7 represent 0,60 fg, 125 fg, 0.25 pg, 0.5 pg, 1 ng and 2.5 ng of
injected eFGF RNA respectively). Animal caps were removed at stage 8,
cultured to stage 10.5 and processed for RNase protection Probing was
carried out for GATA-2,
GA TA-3, Xbra and EF1-a
as a loading
control.
t-
.......
'°
1
(B)
FGF does not induce NCAM but does posteriorize as shown by HoxA 7
expression Caps from the same experiment as (a) but aged to stage 21
probed for HoxA7, NCAM and EF1-o:.This experiment was repeated three
times to verify that the rate of GATA down regulation significantly ex~
ceeded the rate of induction of Xbra (p< 0.05).
........
234
5 6 7 8
_
HoxA7
NCAM
.9- EF1-a
768
E. M. Read et al.
stage 11
~LJI,i
"gilnl bFGF
Xbra .....
GATA-3
.....
GATA-2
.....
EF1-a .....
stage 21
ngIml bFGF
t
WE
4 and a dominant negative 8MP receptor RNA on expression of
gla3 in the zebrafish have been reported by us (Neave at al., 1997).
Thus BMP-4 is implicated in defining the DN boundary of GATA
expression.
Several lines of evidence have indicated that the boundary
between neural and nonneural ectoderm results from the antagonistic action of neural inducers and BMP-4 (Sasai et al., 1995;
Wilson and Hemmati-Brivanlou, 1995). We have previously shown
that the neural inducer noggin plays a role in defining regions of
GATA-2 expression in Xanopusembryos (Walmsley at al., 1994).
Thus noggin was shown to down regulate the expression of GAT A2 in animal cap ectoderm by converting GATA-2 positive animal
cap ectoderm to GATA-2 negative, NCAM positive neural tissue.
Using the same experimental approach, we have now shown that
noggin has a similar effect on GAT A-3 expression (data not shown)
and we have recently reported that Xenopus noggin suppresses
ectodermal gla3 expression in zebrafish (Neave at al., 1997).
Thus, in both frogs and fish, the neurallnonneural
boundary of
GAT A expression may reflect the antagonism between the activities of BMP-4 and a dorsalizing activity such as noggin.
. - HoxA7
-.
....
:II
.... 1-.1'
_ _ 4
-
1 2
II
NCAM
8-EF1-a
3 4
5 6
7 8 9 10 11 12
Fig. 6. FGF posteriorizes
gastrula
nonneural
ectoderm
and down
regulates GAT A-2 and 3 in the absence of mesoderm
and neural
induction. RNA from ventral gastrula caps aged to stage 10.5 then treated
with bFGF in the range 0-100 ng/ml and incubated to stage 11 (lanes 1-4)
or stage 21 (lanes 7-10) was probed by ribonuclease
protection
for
expression of GA TA-2, GATA-3 and Xbra (fanes 1-6), NCAM (doublet) and
HoxA7 (lanes 7-12) and EF1-a as a loadingcontrol. WE, whole embryo RNA
at stage 1/ (lane 5) and stage 21 (lane 11). Negative controls were 20 tlg
tRNA (t, lanes 6 and 12). HoxA7 up-regulation
was confirmed by scanning
densitometrv
taking In to account the loading controls.
ectoderm, being excluded from neural ectoderm, neural crest and
pia codal precursors. This expression is strongest anteriorly, reducing to zero posteriorly. Thus both the presence and patterning of
GATA-2 and 3 expression in nonneural ectoderm has been
evolutionarily conserved between frogs and fish. Figure 10 depicts
diagrammatically
the conserved
boundaries ot GATA-2 and 3
expression in nonneural ectoderm of frogs and fish at gastrula
stages.
Creating the DN boundary: inhibition of BMP-4 signaling
abolishes the expression of GA TA-2 and 3 in Xenopus animal
cap ectoderm
BMP-4 has recently been shown to playa vital role in defining
ectodermal fate. To ask whether BMP-4 might be involved in
defining the neurallnonneural
boundary of GATA-2 and 3 expression, we injected RNA encoding a dominant negative truncated
BMP-4 receptor (tBR) into the animal pole ectoderm of all four
blastomeres at the four cell stage, isolated animal caps at stage 8
and looked for effects on GATA-2 and 3 expression over a time
course of4 h to approximately 14 h (GIN, Fig. 2). Both GATA-2and
3 were down regulated
by the inhibition of BMP-4 signaling
resulting from injection of the truncated
receptor (lanes tBR). The
down regulation
of GAT A-2 can already be detected
in caps
incubated for 5.5 h (longer exposures
show that this is true also for
GAT A-3, which is expressed
at a lower level than GAT A-2, data not
shown). In caps incubated overnight to early neurula when NCAM
expression
starts to be detectable,
the down regulation ot GATA2 and 3 expression is accompanied
by the induction of NCAM
transcription (Fig. 2, doublet in lane tBR of GIN sample) implying
that a conversion of non neural to neural ectoderm has taken place.
These effects were not seen in uninjected embryos nor when a
control mutant FGF receptor, d50 (Amaya et al., 1991), was
injected (Fig. 2, lanes ui and d50). In addition, injection of BMP-4
RNA in tothe animal pole results in stimulation of GATA-2 expression in Xenopus embryos (data not shown). Similar effects of BMP-
Mesoderm induction by FGF and activin down regulates
GA TA.2 and 3 expression in Xenopus and zebrafish embryos
In frog gastrulae, a ring of Xbra expression in the mesoderm
around the blastopore is induced and maintained by FGF signaling
(Isaacs al al., 1994; Schulte-Merker and Smith, 1995). Similarly, in
fish, nt/expression in the germ ring is dependent on FGF (Griffin et
al., 1995). We have shown that, in both frogs and fish, GATA-2 and
3 expression is excluded from the blastoporal mesoderm and the
germ ring (Fig. 1). Thus FGF is present at the right time and place
to playa role in setting up the GATA expression boundary between
nonneural ectoderm and mesoderm.
To test for effects of FGF on the expression of GATA-2 and 3 we
injected eFGF RNA in to Xenopus and zebrafish embryos and
monitored GAT A expression by whole-mount in situ hybridization.
In both species GATA-2 and 3 were strongly down regulated by
eFGF (Fig. 3a-h). When zebrafish embryos were injected with
XFD, a dominant negative receptor for FGF (Amaya at ai, 1991),
and doubly stained for gla2 (purple) and nil (red), ectopic gta2
expression extended posteriorly into the ventral blastoderm margin where nil expression was eliminated (Fig. 3i). However, gla2
expression was not seen where XFD abolished ntl expression
more dorsally. The normal expression of nt/and gta2in uninjected
embryos is shown in Figure 3j. Similar results were found for XFD
injected embryos probed for gla3 and nil (data not shown). Taken
together these data indicate that FGF signaling is required to
exclude GAT A expression from the germ ring and that the dorsoventral patterning of GAT A factors is maintained even during
ectopic expression in the germ ring.
Confirmation of the loss of GATA expression in animal pole
ectoderm resulting from eFGF injection of whole embryos was
obtained in isolated Xanopas animal caps where both GATA-2 and
3 were downregulated by injection of eFGF RNA or exposure to
bFGF protein (see Fig. 5Aand data not shown). The downregulation
of GATA-2 and 3 in caps was accompanied
by induction of
mesoderm as measured by Xbra expression. Similar results were
obtained when animal caps were treated with activin (Walmsley et
al., 1994 and data not shown). Furthermore, in XFD injected caps,
GATA expression levels were undisturbed and treatment of XFD
injected caps with activin, where mesoderm was not induced, also
FGF and Wnts pattern J1(JI1neuralectoderm
had no effect on GATA levels (data not shown). Taken together
these experiments show that when FGF is absent or low in
ectoderm, GATA expression is permitted, but raising the levels of
FGF or of activin results in down regulation of GATA expression
associated with the conversion of ectoderm to mesoderm. Therefore both FGF and FGF-dependent activin-like activities could be
playing roles in setting up the boundary of GATA expression
between nonneural ectoderm and mesoderm.
~
Xbra.....
RA activity is not consistent
with a role in patterning
GA TA
factor expression
in the nonneural
ectoderm
RA has been demonstrated to act as a posteriorizing molecule
during gastrula stages. Measurement of the levels of activity of
endogenous retinoids in Xenopus embryos has revealed that their
distribution is complex but may be present in a gradient in the
anteroposterior
axis, being high in the posterior and low in the
anterior regions of the embryo (Chen ef a/., 1994). We therefore
asked if RA might playa role in patterning GATA-2 and 3 expression in the nonneural ectoderm along this axis. Animal cap ectoderm
from stage 8~9 Xenopus embryos was treated with 1a-7M RA until
stage 11, the period during which ectodermal patterning is taking
place. The caps were then harvested at stages spanning optimal
GAT A-2 and 3 expression, and monitored by RNase protection. RA
had little effect on the expression of GAT A-2 but up-regulated the
expression of GAT A-3 at all stages tested (Fig. 4). The differential
responses of the two GAT A factors to RA treatment suggests that
RA cannot be responsible forthe graded anteroposterior patterning
of GATA-2 and 3 in non neural ectoderm. In the same experiment
the expression of epidermal keratin was clearly down regulated
(Fig. 4), indicating a possible change in cell fate. However, no such
activity has been attributed to RA and, in this case, the down
regulation of epidermal keratin, a marker of terminally differentiated epidermis, is more likely to represent an inhibition of epider~
mal differentiation by RA (Fuchs and Green, 1981). A role for GAT A
factors in maintaining the undifferentiated state has already been
suggested (Briegel ef a/., 1993; Gave et al., 1997) and in this
instance it seems likely that the responses of GAT A-2 and 3 to RA
are reflecting a block to differentiation rather than NP patterning.
FGF posteriorizes
blastula nonneural ectoderm and contributes to the patterning of GA TA factors within this layer
FGF can posteriorize anterior neural tissue (Cox and Hemmati~
Brivanlou, 1995; Lamb and Harland, 1995). During gastrulation,
FGF expression is located in the posterior regions of Xenopus
embryos (Isaacs ef al., 1992,1994; Isaacs, 1997) and is therefore
present at the right time and place to posteriorize nonneural
ectoderm during the period when posteriorization and patterning of
the neural plate is occurring. To investigate the effects of FGF more
~pg
Wnt3A
WE
.
EF1a..-
flit". i
1 234
HoxA7
It
a....
::
EF1o.
....
5
WE
"
I
..1. .
6 7 8 9
stage 20
pg Wnt3A
...
GATA3 ....
GATA2
Patterning GA TA-2 and 3 expression in the AlP axis
We next asked which signaling molecules might be responsible
for refining the pattern of GAT A-2 and 3 in ventral nonneural
ectoderm along the AlP axis. In both Xenopus and zebrafish
embryos, by early gastrula, expression of GAT A-2 extends more
posteriorly than that of GAT A-3 and the most posterior ectoderm is
GATA free (Fig. 1i-I). This suggested the existence of non-axial
posteriorizing signals which might be patterning the nonneural
ectoderm in asimilarmannerto
that described for neural patterning
in Xenopus and zebrafish embryos (Bang et a/., 1997; Woo and
Fraser, 1997).
stage 13
stage 10.5
pg Wnt3A
769
10 t
stage 20
~pg
WE
!
Wnt3A
WE
8]
Actin ....
R:
Cyto
Actin
!I=' I.
!
NCAM..-
EF1a-.1',._.~'
11 121314
.~
15
t
..-
16 17 18 19
20 t
Fig. 7. XWnt3A posteriorizes
blastula nonneural ectoderm and down
regulates GAT A factor expression
in the absence of mesoderm
and
neural induction.
50, 100and 150pgXWnt3A RNA was injected in to the
middle of the animal pole of single cell stage Xenopus
embryos.
Animal
caps from uninjected and injected embryos
were removed at stage 8 and
were cultured in 1xMBS to stage 10.5, 13 or stage 20. RNA prepared from
stage 10.5 caps was probed by ribonuclease protection for Xbra and EF1a. RNA from stage 13 caps was probed for GATA-2 and 3 and EFla and
RNA from stage 20 caps was probed for HoxA7, NCAM and muscle actin.
EF1-a and cytoskeletal actin were used as loading controls. Whole embryo
RNA (lanes 5, 10 15 and 20) and tRNA (fanes marked t) were used as
positive
and negative
controls
respectively.
HoxA7 up-regulation
was
verified by scanning densitometry
taking into account the loading controls
Bands in the lane adjacent to fane 15 were due to leakage from lane 15
closely, we injected eFGF RNA over a wide concentration
range
and monitored expression
in animal caps of GAT A-2, GAT A-3 and
Xbra at stage 10.5 when Xbra expression,
and therefore mesoderm, is first induced. It was important to monitor GAT A expression
at the same time as Xbra to ensure that any effects on GAT A
expression
were not downstream
of mesoderm
induction. In addi~
tion we monitored expression of NCAM and HoxA7 at stage 21,
when both are easily detectable, as a measure of neural induction
and posteriorization respectively. We did not attempt to measure
GATA expression at this late stage since ectodermal GATA expression has declined to low levels by late neurula (Walmsley etal.,
1994).
Induction of Xbra at low levels of FGF was extremely weak yet
strong down regulation of GATA-2 and 3 was observed (Fig. 5A).
The weakness of the Xbra signals at [ow FGF levels indicates that
very few ectodermal celis within the cap had been converted to
mesoderm, so the strong reduction in GATA expression suggests
770
A
E. M. Read et al.
An
GASTRULA
ECTO
DERM
v
D
1
]
[
=GATA-2
A:-IO GATA-3
II =GATA-2 ONLY
MESEI\"DO
DERM
J
Veg
YOLK I
ENDOD
ERM
Hl=GATAfREE
B
NEURULA
DORSAL
NEURAL PLATE
POSTERIOR
ASTERIOR
VE:"ITRAL
Fig. 8. A model
for the
patterning
of GATA-2
and
GATA-3
in the
nonneural
ectoderm.
(A and B). Diagrammatic
representation
of GATA
expression in early gastrula and neurula embryos. Zones 1-/1/represent
the
distinct areas of GATA expression
within the AlP axis {which at blastula/
gastrula stages is roughly equivalent to the animal/vegetal axis} of the
nonneural ectoderm of frogs and fish as determined by whole-mount in situ
hybridization. Patterning of GA TA expression
in to zones I-III by posteriorizing
agents such as FGF and/or Wnts is set up dunng fate blastula/early gastrula
(A). Subsequent
convergent extension movements
lead to the pattern
seen
In
early neurula embryos
as demonstrated
for Xenopus
(B).
that FGF has had an effect on ectoderm independent of mesoderm
induction. Evidence for posteriorizing
activity of FGF in these
experiments came from expression of the posterior gene HoxA7
which was strongly up regulated (Fig. 5B). As expected for blastula
caps (Isaacs et al., 1994), this posteriorization
occurred in the
absence of neural induction measured by NCAM expression (Fig.
5B). So the down regulation of GATA-2 and 3 does nof represent
conversion of nonneural to neural ectoderm. These results suggest that, at very low levels, eFGF can posteriorize non neural
ectoderm in the absence of mesoderm and neural induction, and
that down regulation of GAT A factor expression is a consequence
of this posteriorization.
FGF posteriorizes
gastrula non neural ectoderm and influences GA TA factor pattern within this layer
As an alternative way of studying the effects of FGF on GATA
expression in the absence of mesoderm induction, presumptive
epidermal cells were exposed to FGF after sfage 10.5 when they
are no longer competent to make mesoderm in response to FGF
(Jones and Woodland, 1987). Caps were explanted at stage 10.5.
biasing the eKplanffowards the venfral side, and treafed with bFGF
over a concentration range of 0-100 ng/ml.
We found that GATA-2 and 3 levels in gastrula ectoderm were
down regulated by bFGF in a concentration-dependent
manner in
the absence of mesoderm induction, as revealed by the lack of
Xbraexpression
(Fig. 6, lanes 1-4). The down regulation. however.
was relatively modest (approximately
two-fold). Evidence for
posteriorization in these experiments came from expression of the
posterior gene HoxA7which was up regulafed (allowing for loading
differences and verified by scanning densitometry -Fig. 6, lanes 710). Posteriorization occurred inthe absence of neural induction as
measured by NCAMexpression
(Fig. 6. lanes 7-10). Therefore the
down regulation of GAT A-2 and 3 does not result from conversion
of nonneural to neural ectoderm. These results demonstrate that
FGF can posteriorize nonneural ectoderm from gastrulae in the
absence of mesoderm
and neural induction.
However, this
posteriorization
is accompanied
by a relatively modest down
regulation of GATA factor expression raising the possibility that
FGF functions predominantly during blastula stages and/or that
another signal is normally involved.
XWnt3A posteriorizes
blastula non neural ectoderm and contributes to the patterning of GA TA factors within this layer
Since the down-regulation
of GAT A expression by FGF in
gastrula nonneural ectoderm was incomplete we looked for further
posteriorizing activities which might contribute to patterning of the
GATA factors within the nonneural ectoderm. Both XWnt3A and
XWntB have been reported to posteriorize anterior neural tissue
(McGrew elal., 1995; Fredieu elal.1997).
We injected XWnt3A
RNA over a concentration range, into the middle of the animal cap
of single cell embryos, dissected out caps at stage 8 and cultured
them to appropriate stages for molecular analysis by RNase
protection. Posteriorization of ectoderm by XWnt3A was indicated
by the induction of HoxA7in Wntinjected caps analyzed at stage
20 (taking into account loading differences and confirmed by
scanning densitometry-Fig.
7, lanes 11-15), confirming a previous
report (McGrew. el ai, 1995). Unlike FGF. Wnt3A does not induce
mesoderm in animal cap ectoderm (confirmed below), we therefore analyzed GATA-2 and 3 around the peak of their ectodermal
expression (stage 13). At this stage. both GAT A 2 and 3 were down
regulated but only atthe highest concentration of Wnlused (Fig. 7.
lanes 6-10). indicating that HoxA7 is more sensitive to the Wnl
signal than GATA-2 or 3. As reported previously (McGrew et al..
1995), XWnt3A expression in caps did not induce neural tissue
(indicated by NCAM expression. Fig. 7. lanes 11-15) nor did it
induce dorsal mesoderm indicated by monitoring actin expression
(Fig. 7. lanes 16-20). In addition, XWnt3A failed to induce Xbra
expression in stage 10.5 caps (Fig. 7. lanes 1-5), indicating the
absence of any mesoderm induction. Thus we have demonstrated
that a Wnt family member can posteriorize nonneural ectoderm
and down regulate the expression of GAT A factors in this tissue in
the absence of mesoderm and neural induction.
Discussion
We have described the early boundaries of expression of the
transcription factors GATA-2 and 3 (gta2 and 3) in the nonneural
ectoderm of frogs and fish and found them to be highly conserved
between the two species. Two clear boundaries were obseNed,
one between neural and nonneural ectoderm, the other between
FCF and Wnts pattern nonneural
nonneural
ectoderm and mesoderm. In addition, along the AlP
axis, expression of the two GATA factors is graded and distinct
limits of expression are seen, GATA-2 being expressed more
posteriorly than GATA-3 such that, by early neurula, three zones
of distinct GATA patterning can be seen representing head, trunk
and tail ectoderm (Fig. 8B). The possible functional significance of
this distinctive patterning is suggested as follows. Zone I represents the ectoderm overlying the presumptive branchial arches. A
role for ectoderm in specifying skeletal pattern in the head has been
suggested (Thorogood, 1988) and the importance of GATA-3 in
this region is suggested by the phenotype of the GAT A-3 homozygous mutant mouse which commonly exhibits disruption of
facial structures (Pandolfi eta/., 1995). In zone II, dorsa-laterally,
GAT A-2 positive ectoderm covers prospective somite and dorsolateral plate (the region which includes future definitive blood,
endothelial and nephric systems). Recent evidence in the axolotl
embryo has demonstrated
a role for ectoderm from this region in
signaling to underlying mesoderm for the directional guidance of
migrating pronephric duct and in somite fissure formation (Drawbridge et al., 1995). Ventrally, GATA-2 positive ectoderm represents the superiicial ectoderm of the future blood islands. Ventral
ectoderm has been shown to stimulate hematopoietic
differentiation (Maeno et al., 1992) and GAT A-2 has been strongly implicated
in this ectoderm/mesoderm
interaction (Maeno el al., 1996). Finally, the GATA free region of the nonneural ectoderm (zone III)
represents
the ectodermal
component of the future tail as defined
in fate mapping studies in Xenopus (Tucker and Slack, 1995),
suggesting that these GAT A factors play no role in the activity of tail
forming genes. These broad regions of transcription factor expression may represent the beginnings of functional segmentation
of
nonneural ectoderm along the AlP axis as suggested
in the
concept of the 'ectomere' (Couly and Le Douarin, 1990).
The dorsoventral
axis
Our results suggest that the neural/nonneural
boundary of
expression of GAT A factors in ectoderm is set up by the antagonistic action of dorsalizing activities such as noggin with BMP-4
(Walmsley et ai, 1994; Neave et al., 1997; this study). Since BMP2
and 7 also interact with noggin and with chordin in a similar fashion
to BMP-4 (Piccolo et al., 1996; Zimmerman et al., 1996) it is likely
that some redundancy exists between the BMP family and organizer-associated
dorsalizing molecules and that all of these molecules could play roles in restricting the expression of GATA
factors to the non neural ectoderm. In addition to representing the
limits of the neural plate, this boundary has an important functional
role. It is clear that the interaction of nonneural ectoderm with
neural ectoderm is responsible for the formation of the neural folds
and the neural plate border which contains both neural crest and
placodal tissues (Maury and Jacobson, 1989; Dickinson et al.,
1995). GA TA-2 and 3 may have important functions in defining this
border.
A homolog of GATA-2 and 3 is the pannier gene in Drosophila.
ectoderm
771
neural/nonneural
decisions within the imaginal disc. In the embryo
proper, pannier is expressed in the dorsal ectoderm of Drosophila
(Winick et al., 1993) which, given the inversion of the dorsoventral
axis between vertebrates and invertebrates (Holley et al.., 1995), is
the equivalent location to GATA-2 and 3 in frogs and fish. Very
recently, a C. elegansGAT A homolog, Ell-I, has been shown to be
absolutely required for epidermal cell fate (Page et al., 1997). It
therefore seems that GATA function as well as expression pattern
in embryonic ectoderm has been highly conserved during evolu-
tion.
The anteroposterior
axis
Creation of the boundary between nonneural ectoderm and
mesoderm at late blastula stages seems to involve FGF and/or
activin-like molecules. Both these molecules down regulate the
GATA factors in ectoderm when present at concentrations sufficiently high to induce mesoderm. Therefore these molecules are
likely to be involved in the suppression of GATA-2 and 3 in the
region of the non neural ectoderm around the blastopore. This
expression pattern is refined at some point between late blastula
and early neurula, such that the two factors become graded in the
AlP axis, high anteriorly and low or absent in more posterior
regions. An expression pattern such as this might well result from
the activity of signals produced in posterior regions of the gastrula
embryo acting negatively on GATA expression equivalent to the
non-axial signals demonstrated in Xenopus and zebrafish to
posteriorize neural ectoderm (Bang et al., 1997; Woo and Fraser,
1997).
RA has been shown to be a potent posteriorizer of the neural
plate during these stages but does not appear to be the agent that
down regulates the GATAfactors in nonneural ectoderm, since
it
has little effect on GATA-2 and strongly up regulates GATA-3. In
this assay the expression of epidermal keratin (a marker of terminal
epidermal differentiation) was shown to be repressed, indicating
that differentiation of the ectoderm had been prevented. There is
no
evidence that RAtreatmentof animalcap ectoderm produces
a change in cell fate, therefore the response of the two GATA
factors is more likely to reflect a change in their expression levels
at different stages of differentiation of the epidermal precursors.
Although a role for FGF in the induction and posteriorization of
neural ectoderm has been claimed (Lamb et al., 1993; Cox and
Hemmati-Brivanlou, 1995; Kengaku and Okamoto, 1995), more
recently Kroll and Amaya have shown that signaling through the
FGF receptor is not required for neural induction or antero-poste-
rior patterning in transgenic embryos (Kroll and Amaya, 1996).
These contradictory results may be explained by a multiplicity of
FGFreceptors or alternative neural inductionand posteriorization
pathways employed in XFD transgenic embryos. In addition, Woo
and Frazerwhen examining candidates for the non-axial signaling
which patterns neuraltissue of zebrafish embryos in the AlP axis
concluded that FGF alone cannot fulfil this role (Woo and Fraser,
1997). In the Xenopus study too, Bang et. al. concluded that neither
The pannier gene codes for a zinc finger protein with strong
RA nor FGF acting alone
homology in the finger region to the zinc fingers of GATA factors
(Ramain et aI., 1993). Pannier acts as a repressor of achael.e and
scute, transcription factors expressed at sites where neural precur-
activity. Here, we have provided evidence that FGF can posteriorize
Xenopus nonneural ectoderm, as measured by the up-regulation
of the posterior gene HoxA7, in the absence of mesoderm and
neural induction. This activity of FGF appears to contribute to the
down regulation of GATA expression in posterior regions of the
sors develop in the imaginal disc. Pannier mutants with lesions in
the zinc finger domain display an overexpression
of achaete and
scute and the development of extra neural precursors. This Drosophila GATA homolog, therefore, plays an important role in
couldrepresent the non-axialposteriorizing
developing embryo. From our data, the effecf is stronger when the
cells are exposed at late blastula stages (stage 8-10.5) and
----
772
E. M. Read et al.
decreases during gastrulation
to about two-fold between stages
10.5 and 11. Thus in Xenopus embryos, much of the patterning of
GATA factors within the ectoderm by FGF may have occurred
before gastrulation. This would account for the relatively modest
effect of FGF on GATA expression in gastrula caps. An additional
possibility is that another factor is involved.
A candidate for such an additional factor is a member of the Wnt
family. We demonstrate
here that XWnt3A can posteriorize
nonneural ectoderm and, at high concentration, can down regulate
GATA factor expression in this tissue. Thus XWnt3A acts to
posteriorize
all ectoderm.
Both XWnt3A
and XWntBare
expressed
in posterior ventra-lateral
mesoderm during gastrulation in nonaxial regions around the blastopore. So a number of Wnt family
members may contribute to posteriorizing and patterning ectoderm
at these early developmental
stages. Our experiments do not
address whether Wntsact directly on ectoderm to posteriorize it or
whether Wnts trigger production of a dominant posteriorizing
morphogen which then patterns ectoderm. Precedence for Wnts
activating a secondary morphogen which then acts over a distance
to alter cell fate has been reported for other vertebrate and
invertebrate systems (see Fredieu et al., 1997 and references
therein ).
The positive and negative effects of BMP-4 and posteriorizing
agents respectively on GAT A-2 and 3 suggest the following model
for establishing the observed expression patterns (Fig. 8). At or just
before the beginning of gastrulation, ectodermal GAT A expression
is activated by BMP-4 in regions beyond the influence of mesoderm induction and dorsalizing signals from the organizer (zones
I and II, Fig. 8A). FGF and Wnts, produced initially inthe mesoderm
and maintained in posterior regions as gastrulation proceeds,
down regulate GA TA expression in proximal nonneural ectoderm.
The most posterior ectoderm which experiences the longest exposure to posteriorizing
signals during epiboly and gastrulation
becomes GATA free (Zone ill, Fig. 8A) whilst in more anterior
regions GATA expression is permitted (Zones I and iI). The
differential extents of expression of GAT A-2 and 3 in the anteroposterior axis could additionally reflect graded suppression starting from different maximal activation levels (Figs. 2-7 -expression
levels of GAT A-2 are always higher than GAT A-3). At low levels of
FGF and Wnts both GATA-2 and 3 are expressed (zone I), at
intermediate levels only GAT A-2 is seen (zone iI) and althe highest
levels (zonelll) neither GATA factor can be detected.
Materials
and Methods
Embryos and dissections
Production and dissections of Xenopus and zebrafish embryos were as
described (Walmsley et al., 1994; Neave et al., 1995). For blastula caps, the
top one third of the animal pole ectoderm was removed at stage 8-8.5 and
cultured to appropriate stages for molecular analysis. Gastrula cap ectoderm
was dissected at stage 10.5 biasing the dissection towards the ventral side.
Caps were then transferred to 1xMBS containing 0.5% BSA and 0, 20, 50 or
(Wright et aI., 1990; Godsave
(Mayor et aI., 1995).
et af., 1994), Xbra (Smith et al., 1991), slug
Zebrafish whole-mount in situ hybridizations were pertormed as described (Neave et al., 1995) and double whole-mount in situ hybridizations
were periormed as in (Hauptman and Gerster, 1994). The gta2 probe used
was transcribed from a fun length cDNA cloned from a zebrafish neurula
library (kind gift of David Grunwald) using zinc fingers of Xenopus GAT A-2a
as probe. Other probes were: gta3 (Neave et al., 1995), ntl (Schulte-Merker
etaf., 1992), dlx3 (Akimenko etal., 1994), pax2 (Krauss etal., 1991).
RNA preparation and ribonuclease protection assays
Preparation of RNA and ribonuclease protection assays were as described (Walmsley et a/., 1994). The protection probe for XGATA.3 was
prepared from a 150bp Mspl fragment from the 5' end of GAT A-3 cDNA
inserted in to the Cia! site of pGEM7. Transcription
with T7 polymerase
gives antisense RNA. Other probes have been described: GATA-2, NCAM
and EF-1a (Walmsley et a/., 1994), Xbra (Smith et al., 1991), HoxA7
(Condie and Harland, 1987).
Preparation of in vitro transcribed RNA for injection
In vitro transcribed RNA was prepared using a Megascript kit (Ambion)
as per the manufacturer's
instructions except that capped GTP was used
at a lower concentration of 5 mM. For Xenopus embryos, RNA was injected
in a volume of 4 nl in to the centre of the animal pole ectoderm at the single
cell stage. eFGF RNA was supplemented with carrier tRNA (10 pg/nl final)
to avoid losses at the low end of the titration range. For zebrafish embryos,
RNA was prepared as in (Cornell and Kimelman, 1994) using Megascript
kit transcribing for 4 h with 4:1 cap:GTP for eFGF and 2:1 cap:GTP for XFD
and d5D. RNA was injected in a volume of approximately 20Dpl into one cell
of a 1-4 cell stage embryo.
Signaling molecules
Human recombinant bFGF was purchased from Gibco. Activin was a
tissue culture supernatant (gift from Jim Smith, NIMR, Mill Hill). Noggin was
expressed post MBT by injection of the plasmid pCSKA nag and into the
animal cap of single cell or four cell stage embryos respectively (Walmsley
et al., 1994). Templates for transcription of in vitro RNA for eFGF, the
dominant negative BMP-4 and FGF receptors, XWnt3A and d50, a mutant
FGF receptor control, were prepared as described (Amaya et al., 1991;
Isaacs et al., 1994; McGrew et al., 1995; Northrop et al., 1995). All-trans RA
was purchased from Sigma.
Acknowledgments
We thank Richard Harland, Mike Sargent, David Wilkinson and M.
Pannese for gifts of whole-mount probes; Lynn McGrew for the Wnt3A
construct and Kate Kirwan for excelfent photographic assistance. Particular thanks go to the members of Matthew Guille's group who went beyond
the calf of duty to help with bFGF problems. This work was supported by the
MRC and by studentships
from the BBSRC (EM.R)
and the Weflcome
Trust (B.N.)
References
AKIMENKO,
M.-A., EKKER, M., WEGNER. J., UN, W. and WESTERFIELD,
M.
(1994). Combinatorial
expression of three zebrafish genes related to Distal-Less:
part of a homeobox
gene code for the head. J. Neurosci.
14: 3475-3486
100 ng/ml bFGF and cultured to stage 11 for measurement
of GATA-2 and
3 and Xbra RNA or stage 21 for measurement of NCAM and HoxA7 RNA.
AMAYA, E., MUSCI, T.J. and KIRSCHNER,
M.W. (1991). Expression of a Dominanf
Negative Mutant of the FGF Receptor Disrupts Mesoderm Formation in Xenopus
Embryos. Ge1/66: 257-270.
Whole-mount
in situ hybridization
For Xenopus, double and single whole-mount in situhybridizations
were
carried out as described (Bertwistle et a/., 1996). For details of DIG and
fluorescein labeled probes see:
GATA-2 (Walmsley et al., 1994), GATA-3 (Bertwistle et al., 1996),
Otx2 (Pannese et al., 1995), XKrox20 (Bradley et aI., 1992), HoxB9
BANG, A.G., PAPALOPULU.
N.. KINTNER,
C. and M.D., G. (1997). Expression
of
Pax-3 is initiated in the early neural plate by posteriorizing signals produced by the
organizer and by posterior non-axial mesoderm. Development
124: 2075-2085.
BERTWISTLE,
D., WALMSLEY. M.E.. READ, E.M., PIZZEY, JA and PATIENT, R.K.
(1996). GATA factors and the origins of adult and embryonic blood in Xenopus:
responses
to retinoic acid. Mech. Dev. 57: 199-214.
FGF and Wnts pattern nonnellral ectoderm
BRADLEY,
L.C., SNAPE, A., BHATT, S. and WILKINSON,
and expression
of expression
of the Xenopus
in rhombomeres
D.G. (1992). Thestructure
Krox-20 gene: conserved
and divergent
ISAACS, H.Y., POWNALL, M.E. and SLACK, J.M.W. (1994). eFGF regulates
patterns
and neural crest. Mech. Dev. 40: 73-84.
CHEN,
Y., HUANG,
Retinoids
differentiation
in a hormone
L. and SOLURSH,
in the Early Xenopus
dependent
M. (1994).
laevis Embryo.
Genes Dev.
manner.
A Concentration
Gradient
of
Dev. BioI. 161:70-76.
mesoderm
of Xenopus.
Genes Dev. 7: 13-28.
B.G. and HARLAND,
R.M. (1987).
Posterior
CORNELL, RA and KIMELMAN, D. (1994). Activin
requires FGF. Development
120: 453-462.
chimeras: evidence for a segmental pattern
neuromeres,
Development
108: 543-558
tissue recombination
mediated
induction
avian
corresponding
A (1995). Caudalisation
to
of neural fate by
121: 4349-4358
CREECH KRAFT, J., SCHUH, T., JUCHAU, M. and KIMELMEN,
D, (1994). The
retinoid X receptor ligand,9~cis-retinoic
acid, is a potential regulator of early
Xenopus development.
Proc. Nat!. Acad. Sci. USA 91: 3067-3071.
DICKINSON,
M.E., SELLECK,
M. (1995). Dorsalisation
MAJ.,
MCMAHON,
AP. and BRONNER-FRAZER,
of the neural tube by the nonneural
ectoderm.
Develop-
ment 121: 2099-2106.
DRAWBRIDGE,
J., WOLFE,
The epidermis
AE.,
DELGADO,
neural pattern in Xenopus
by planar
Y.L. and STEINBERG,
M.S. (1995).
information
pronephric
VRIES,
causes
system.
an anteroposterior
transformation
Nature 340: 140-144.
in the developing
central
DE
Acid
nervous
J.R., Cui, Y., MAIER, D., DANILCHIK,
E. and GREEN,
human
H. (1981). Regulation
M.V. and CHRISTIAN,
J.L. (1997).
GODSAVE,
of terminal
differentiation
S.F., DEKKER,
E.J., HOLLING,
GOVE,
C.,
of the mesoderm.
WALMSLEY,
PARTINGTON,
T., PANNESE,
M., BONCINELLI,
E. and
Dev. Bioi. 166: 465-476.
M.,
G., BOMFORD,
GATA-6 in Xenopus embryos
16: 355-368.
GRAFF, J.M., THIES,
NIJJAR,
S.,
BERTWISTLE,
A. and PATIENT,
R. (1997).
blocks differentiation
R,S., SONG, J.J., CELESTE,
D.,
GUILLE,
M.,
Over-expression
01 heart precursors.
A.J. and MELTON,
Studies with a Xenopus BMP receptor suggest that ventral
signals override dorsal signals in vivo, Gel! 79: 169-179.
of
EMBOJ.
reveals seperate
121: 2983-2994.
HAUPTMAN, G. and GERSTER, T. (1994). Two-colour
zation to vertebrate
and Drosophila
D,A. (1994).
HOLLEY,
SA,
JACKSON,
F.M. and FERGUSON,
P.D., SASAI,
embryos.
ectoderm
Trends
mechanisms
in
for formation
whole-mount
Genel.
in sItu hybridi-
10: 266
involving
New perspectives
leads to direct neural
E., HOFFMAN,
system fordorsal-ventral
sog and chordin.
Nature
induction.
patterning
376: 249-252.
on the role of the fibroblast growth factor
family in amphibian development. Get!. Mol. Ufe Sci. 53: 350~361.
ISAACS,
anteroposterior
growth
of cultured
factor
induces
ectoderm
rnorphogen
axis of the central nervous system. Development
cells
for the
121: 3121-3130.
KOLM, P.J. and SIVE, H.L. (1995). Regulation of the Xenopus labialhomeodomain
genes, HoxA 1 and HoxD1 : Activation by retinoids and peptide growth factors. Dev.
KRAUSS, S., JOHANSEN, T., KORZH, V. and FJOSE, A. (1991). Expression of the
zebrafish paired box gene pax[zt-bjduring
1193-1206.
KROLL,
K.L. and
AMAYA,
early neurogenesis.
Transgenic
E. (1996).
nuclear transplantations
reveal FGF signaling
Development 122: 3173-3183.
Xenopus
Development
embryos
requirements
113'
from sperm
during gastrulation.
LAMB, T.M. and HARLAND, A.M. (1995). Fibroblast growth factor is a direct neural
inducer, which combined with noggin generates anterior-posterior
neural pattern
Development
121:3627-3636.
LAMB, T.M., KNECHT,
A.K., SMITH,
N., YANCOPOLOUS,
tion by the secreted
W.C., STACHEL,
S,E.,
G.D. and HARLAND,
polypeptide,
noggin.
Science
cated FGF receptor blocks neural induction
Development
122: 869-880.
TREMML,
G., ROELlNK,
ECONOMIDES,
Neural
R.M. (1993).
A.N.,
induc-
262: 713-718
LOFBERG,
by endogenous
H. and JESSEL,
tiation of neural plate cells induced
ectoderm. Cet/82: 969-979.
by BMp.
Xenopus
Dorsal differen-
T.M. (1995).
mediated
inducers.
signals
from epidermal
J., NYNAS-MCCOY,
A., OLSSON,
C., JONSSON,
L. and PERRIS,
01 initial neural crest cell migration in the axolotl embryo
tissue grafts and extracellular
matrix transplanted
A.G.S. (1988). Spatial organisation
crest cells in the initiation
(Suppl.): 155-169.
on microcarriers.
of the epithelium
of the mammalian
LUMSDEN, A AND KRUMLAUF, R. (1996).
Science 274: 1109-1115.
R.
by
Dev. BioI. 107:
tooth
Patterning
and the role of neural
germ.
Development
the vertebrate
103
neuraxis.
MAE NO, M., MEAD, P.E., KELLEY, C., KUNG, H.-F., SUZUKI, A., UENO, N. and
ZON, L.I. (1996). The role of BMP-4 and GATA-2 in the induction and differentiation of hematopoietic
MAENO,
MAYOR,
mesoderm
in Xenopus
laevis. Blood 88: 1965-1972.
M., ONG, R.C. and KUNG, H. (1992). Positive and Negative
R., MORGAN,
R. and SARGENT,
for Erythrocytes
M.G. (1995).
MCGREW, L.L., HOPPLER, S. and MOON, RT (1997).
cooperatively
pattern antero.posterior
neural ectoderm
Regulation
in Xenopus
Induction
of
laevis. Dev.
of prospective
Wnt and FGF pathways
in Xenopus. Mech. Dev.
69: 105-114.
MCGREW,
Y., LU, B., DE ROBERTIS,
E.L. (1995). Aconserved
in insects and vertebrates
H.V. (1997).
Xenopus
Basic fibroblast
neural crest of Xenopus. Development 121: 767-777.
HAWLEY, S.H.B., WUNNENBERG-STAPLETON,
K., HASHIMOTO,
C., LAURENT,
M.N., WATABE, T., BLUMBERG,
BW. and CHO, K.W.Y. (1995), Disruption of
BMP signals in embryonic
Genes Dev. 9: 2923-2935.
H. (1993).
the Differentiation of Ventral Mesoderm
Growth Differ. 34: 567-577.
mesoderm.inducing
GRIFFIN, K., PATIENT, R. and HOLDER, N. (1995). Analysis 01 FGF function
normal and no tail zebrafish embryos
of the trunk and the tail. Developmenl
M. and OKAMOTO,
differentiation
of neural tube and neural crest lineages
from Xenopus gastrula. Development
119: 1067-1078.
LUMSDEN,
of cultured
DURSTON, A (1994). Expression patterns of Hoxb genes inthe Xenopus embryo
suggest roles in anteroposterior
specification of the hindbrain and in dorsoventral
patterning
101: 557-563.
442-459.
by vitamin A, Ge1/25: 617-625.
keratinocytes
of a novel
(1985). Stimulation
Xwnt-8 and Lithium can act upon either Dorsal Mesodermal or Neurectodermal
Cells to Cause a Loss of Forebrain in Xenopus Embryos. Dev. BioI. 186: 100-114
FUCHS,
Expression
inducing
JONES, E.A. and WOODLAND,
H.R. (1987). The development
of animal cap cells in
Xenopus: a measure of the start of competence to form mesoderm. Development
LlEM, K.F.J.,
J,P.M., HAGE, W.J., HENDRIKS,
HFJ.,
A.J" TIMMERMANS,
N,J., HEIDEVELD,
M. and NIEUWKOOP,
P.D. (1989). Retinoic
FREDIEU,
(1992).
A new candidate
LAUNAY, C., FROMENTOUX, V., SHI, D.-L. and BOUCAUT, J.-C. (1996). A trun-
for the migrating
duct in Ambystoma mexlcanum embryos. Dev. Bioi. 172: 440-451.
DURSTEN,
is a source of directional
J.MW.
factor for mesoderm
formation and anteroposterior specification. Development 114: 711-720.
STAHL,
DONIACH. T. (1992).lnduction
of anteroposterior
signals. Development
(Suppl.): 183-193.
D. and SLACK,
embryo.
Xbra
EMBO J. 13: 4469-4481.
Bioi. 167: 34-49.
in embryonic
in the ectoderm
and bFGF. Development
of a homeobox
mesoderm
N.M. (1990). Head morphogenesis
COX, W.G. and HEMMATI-BRIVANLOU,
H.Y., TANNAHILL,
gastrulation.
KENGAKU, M. and OKAMOTO, H. (1995). bFGF as a possible
expression
gene in early Xenopus embryos. Development 101: 93-105
COUL Y, G, and LE DOUARIN,
during Xenopus
FGF in the Xenopus
KENGAKU,
CHRISTIAN, J.L and MOON, R.T. (1993). Interactions between Wnt-8and Spemann
organiser signaling pathways generate dorsoventral
pattern in the embryonic
CONDIE,
expression
ISAACS,
BRIEGEL, K., LlM, K.-C., PLANCK, C., BEUG, H., ENGEL, J.D. and ZENKE, M.
(1993). Ectopic expression of a conditional GATA-2/estrogen
receptor chimera
arrests erythroblast
7: 1097-1109.
773
L.L., LAI, C.-J. and MOON,
terior Neural Axis through
RT
(1995). Specification
of the Anteropos-
Interaction of the Wnt Signaling
with noggin and follistatin. Dev BioI. 172: 337-342.
MaURY,
J,D. and JACOBSON,
boundaries
57.
between
Synergistic
AG.
(1989). Neural fold formation
neural plate and epidermis
Cascade
at newly created
in the axolotl. Dev. Bioi. 133:44-
NEAVE, B., HOLDER, N. and PATIENT, R. (1997). A graded response to BMp-4
spatially coordinates patterning of the mesoderm and ectoderm in the zebrafish.
Mech. Dev. 62: 183-196.
NEAVE, B., RODAWAY, A, WILSON, SW., PATIENT,
R. and HOLDER,
N. (1995).
Expression
of zebrafish GATA 3 (gta3) during gastrulation
and neurulation
suggests a role in the specification of cell fate. Mech. Dev. 51: 169-182.
774
E. M. Read et al.
NORTHROP, J., WOODS. A., SEGER, R., SUZUKI, A., UENO, N., KREBS,
E. and
KIMELMAN, D. (1995). BMP-4 regulates the dorsal-ventral differences in FGFJ
MAPKK-mediated mesoderm induction in Xenopus. Dev. BioI. 172: 242-252.
SIVE, H.L., DRAPER,
PAGE, B.D., ZHANG, W., STEWARD,
K., BLUMENTHAL,
T. and PRIESS, J.R.
(1997). EL T-1, a GATA-like transcription factor, is required for epidermal cell fates
SMITH, J.C. (1989). Mesoderm induction
amphibian development.
Development
in Caenorhabditis
elegans embryos. Genes Dev. 11: 1651-1661.
PANDOLFI, P.P.,
ROTH, ME, KARIS, A., LEONARD, M.W., DZIERZAK, E.,
GROSVELD,
F.G.. ENGEL, J.D. and LINDENBAUM,
M.H. (1995). Targeted
disruption of the GAT A3 gene causes severe abnormalities
in the nervous system
and in fetal liver haematopoiesis.
PANNESE,
M.,
BARSACCHI,
Nature
Genet.
11: 40-44.
POLO, C., ANDREAZZOLl,
M., VIGNALI,
G. and BONCINELLI,
E. (1995). The Xenopus
is a maternal homeobox
gene that demarcates
regions. Development
121:707-720.
PAPALOPULU,
N. and KINTNER,
reveals that anteroposterior
R., KABLAR,
B.,
homologue of Otx2
and specifies
C. (1996). A posteriorizing
anterior
factor,
retinoic
patterning
controls the timing of neuronal
tion in Xenopus neurectoderm. Development 122: 3409-3418.
BMP-4,
to the vertebrate
transcription
factor GAT A-1. Development
of neural
Xenopus.
SCHMIDT,
H. and DE ROBERTIS,
induction
by the Chd and Bmp-4
Nature 376: 333-336.
J.E., SUZUKI,
4 mediates
37-50.
SCHULTE-MERKER,
to Brachyury
A., UENO, N. and KIMELMAN,
dorsal/ventral
patterning
S. and SMITH,
requires
E.M. (1995). Regulation
antagonistic
J.C. (1995).
signals
D. (1995). Localized
in the early Xenopus
FGF signaling.
patterning
Mesoderm
embryo.
BMP-
Dev. Bioi. 169:
formation
in response
Curro Bioi. 5: 62-67.
SCHULTE-MERKER,
S., HO, R.K., HERRMAN, B.G. and NUSSLEIN-VOLDHARO,
C. (1992). The protein product of the zebrafish homolog of the mouse T gene is
expressed in the nuclei of the germ ring and the notochord of the early embryo.
Development
116: 1021-1032.
SCHULTE-MERKER,
S., vAN EEOEN, F.J.M., HALPERN, M.E., KIMMEL, C.B. and
NUSSLEIN-VOLHARD,
C. (1994). no tail (ntl) is the zebrafish homologue of the
mouse
T (Braehyury)
gene. Development
120: 1009-1015.
R.S., YAMAJI,
N., SONG, J.J., WOZNEY,
P. (1988). The developmental
103 (Suppl.): 141-153.
specification
(1995). The Xenopus
in early
J.M., MURAKAMI,
K.
of the Vertebrate
laevistail-forming
skull.
region.
laevis
in the
120: 2519-2529.
P.A. and HEMMATI-BRIVANLOU,
A. (1995).
Induction
of epidermis
and
inhibition of neural fate by Bmp-4. Nature 376: 331-333.
WiNICK, J., ABEL, T., LEONARD, MW., MICHELSON,
A.M., ChARDON-lORIAUX,
I., HOLMGREN,
R.A., MANIATIS, T. and ENGEL, J.D. (1993). A GATA family
transcription
in
factors
105: 665-677.
WALMSLEY, M.E., GUillE,
M.J., BERTWISTLE,
D., SMITH, J.C., PIZZEY,J.A. and
PATIENT, R.K. (1994). Negative control of Xenopus GATA-2 by activin and
noggin with eventual expression
in precursors
of the ventral blood islands.
WILSON,
RUIZ I ALTABA, A. and JESS ELL, T.M. (1991). Retinoic acid modifies the pattern of
cell differentition
in the central nervous system of neurula stage Xenopus embryos. Development
112:945-958.
and mesoderm-inducing
in Xenopus
VON BUBNOFF, A., SCHMIDT, J.E. and KIMELMAN, D. (1995). The Xenopus
homeobox gene Xgbx-2 is an early marker of anteroposterior
patterning
ectoderm. Mech. Dev. 54: 149-160.
Development
119: 1277-1291.
SASAI, T., LU, B., STEINBEISSER,
A., THIES,
TUCKER, A.S. and SLACK, J.M.W.
Developmen/121:
249-262.
P. (1993). pannier, a
encodes a zinc finger
H. (1990). Identifi-
and UENO, N. (1994). A truncated bone morphogenetic
protein receptor affects
dorso-ventral patterning in the early Xenopus embryo. Proe. Nat!. Acad. Sei. USA
91: 10255-10259.
acid,
Cell 86: 589-598.
protein with homology
SUZUKI,
THOROGOOD,
Development
differentia-
R.M. and WEINTRAUB,
period during primary axis formation
SMITH, J.C., PRICE, B,M.J., GREEN, J.B.A., WEIGEL, D. and HERRMANN,
B.G.
(1991). Expression of the Xenopus homologue of Brachyury(T) is an immediateearly response to mesoderm induction. Ce1/67: 79-87.
body
PICCOLO, S., SASAI, Y., LU, B. and DE ROBERTIS,
E.M. (1996). Dorsoventral
patterning in Xenopus: inhibition of ventral signals by direct binding of chordin to
RAMAIN, P., HEITZLER,
P., HAENLlN,
M. and SIMPSON,
negative regulator of achaete and scute in Drosophila,
B.W., HARLAND,
cation of a Retinoic Acid-sensitive
laevis. Genes Dev. 4: 932-942.
factor is expressed
sophila melanogaster.
along the embryonic
119: 1055-1065
dorsoventral
axis in Dro-
Development
WOO, K. and FRASER, SE (1997). Specification
by Nonaxial Signals. Science 277: 254-257.
WRIGHT, C.V., MORITA, EA, WilKIN,
Xenopus
XIHbox6 homeo. protein,
expressed in proliferating neurones.
of the Zebrafish
Nervous
System
D.J. and DE ROBERTIS, E.M. (1990). The
a marker of posterior neural induction, is
Development
109: 225-234.
ZIMMERMAN,
L.B., DE JESUS-ESCOBAR,
J.M. and HARLAND, A.M. (1996). The
Spemann organizer signal noggin binds and inactivates bone morphogenetic
protein 4. GeIlS6: 599-606.
!&Cf>iVf>d:Ala)" 1998
i1ccejJtnl jrH'publication: June
1998